Table 2.

Multivariable binary logistic regression model in total and generalized estimation equation binary logistic regression model in matched myasthenia gravis patients for SARS-CoV-2 infection with regard to immunosuppression status

	Non-matched MG patients (n = 1352a) Nagelkerke R2 = 0.02		Matched MG patients (n = 203 groups, 936 individualsa) Marginal R2 = 0.01
	OR (95% CI) for COVID-19 infection	p	OR (95% CI) for COVID-19 infection	p
No corticosteroids and no immune suppression (reference)	1		1
Only corticosteroids	0.82 (0.41–1.66)	0.589	0.74 (0.35–1.60)	0.445
First line IST (AZT, MMF, MTX, CSA)	0.56 (0.31–1.01)	0.052	0.60 (0.32–1.12)	0.107
Escalation therapy with rituximab	0.59 (0.23–1.56)	0.289	0.85 (0.29–2.51)	0.769
Age in decades	1.00 (0.99–1.02)	0.692	–
Sex male, ref: female	1.10 (0.68–2.09)	0.541	–
Comorbidities present, ref: no comorbidities present	1.19 (0.68–2.09)	0.541	–
MGFA, ref: 0			–
I	0.39 (0.16–0.97)	0.043	–
II A	0.53 (0.22–1.25)	0.146	–
II B	0.50 (0.21–1.18)	0.112	–
III A	0.80 (0.28–2.28)	0.672	–
III B	0.70 (0.25–1.94)	0.491	–
IV A/IV B/V	1.10 (0.36–3.33)	0.868	–

Odds ratios (OR) and 95% confidence interval (CI) for SARS-CoV-2-infected myasthenia gravis (MG) patients from multivariable binary logistic regression (left side) adjusted for age, sex, MGFA score, and thymectomy status (n = 1352, Nagelkerke R Square: 0.02), and from generalized estimation equation (GEE) binary logistic regression model (right side) after propensity score matching (n = 203 matched groups, 936 individuals) without further adjustment

Abbreviations: AZT azathioprine, CI confidence interval, CSA cyclosporine A, GEE generalized estimation equation, IST immunosuppressive therapy, MG myasthenia gravis, MGFA Myasthenia gravis foundation of America classification, MMF mycophenolate mofetil, MTX methotrexate, n number of included patients, OR odds ratio

^aPatients treated with eculizumab (n = 20) or missing covariate data (n = 16) were excluded from matched and non-matched analysis